Chapter 50

Experimental Study of Measurement Errors in 3D-DIC

Due to Out-of-Plane Specimen Rotation

Farzana Yasmeen, Sreehari Rajan, Michael A Sutton, and Hubert W. Schreier

Abstract In practice, out-of-plane motions usually are not avoidable during experiments. Since 2D–DIC measurements
are vulnerable to parasitic deformations due to out of plane specimen motions, three-dimensional digital image correlation
(3D-DIC) oftentimes is employed. The 3D-DIC method is known to be capable of accurate deformation measurements for
specimens subjected to general three-dimensional motions, including out of plane rotations and displacements. As a result,
there has been limited study of the deformation measurements obtained when using 3D-DIC to measure the displacement
and strain fields for a specimen subjected only to out-of-plane rotation. This paper presents experimental results regarding the
effect of out-of-plane rotation on strain measurement using 3D-DIC. Specifically, full-field deformation results are obtained
during rigid body, out-of-plane rotation in the range 40ı  ™ 40ı using a two-camera stereovision system. Results indicate
that (a) the measured normal strain in the foreshortened direction increases in a non-linear manner with rotation angle, (b)
the normal strain along the direction of the rotation axis is essentially zero for all rotation angles and (c) the in-plane shear
strain is small but increases linearly with rotation angle. Results also indicate that the magnitude of the errors in the strain
are a strong function of how the calibration process is performed, with measurement errors exceeding ˙1400 " for what
would normally be considered “small angle” calibration processes.

Keywords Stereo digital image correlation • Rigid body out of plane rotation • Stereo calibration • Strain errors

50.1 Introduction

Digital image correlation (DIC) is an optical non-contact deformation measurement technique capable of measuring full
field deformation on a specimen’s surface [1]. DIC was first originated and devised for planar surfaces undergoing in-plane
deformations in the early 1980s [2] and extended to general deformations of curved or planar surfaces in the early 1990s
[3–5]. Since that time, the DIC technique has grown into one of the most popular measurement method worldwide. In
principle, the method known as 2D-DIC gives accurate results when a nominally planar specimen is subjected to nominally
in-plane mechanical loading. In such cases, the planar surfaces experience in plane deformation. If the single viewing camera
is placed perpendicular to the specimen, then the measured deformations would reflect the true deformations under ideal
conditions. However, as shown previously [6], any out of plane displacement or out of plane rotation of the specimen will
induce in-plane displacement gradients that deleteriously affect the measured strains. In this work, the authors have focused
on the effects of out-of-plane rotation on 3D-DIC deformation measurements using experimental result. This paper presents
details regarding out-of-plane rotation experiments that were performed with a two-camera stereovision system (3D-DIC or
StereoDIC) to investigate the effect of out of plane rotation on 3D-DIC measurements.

F. Yasmeen () • S. Rajan • M.A. Sutton

Department of Mechanical Engineering, University of South Carolina, Columbia, SC 29208, USA
e-mail: fyasmeen@email.sc.edu
H.W. Schreier
Correlated Solutions Incorporated, 121 Dutchman Blvd, Irmo, SC 29063, USA

© The Society for Experimental Mechanics, Inc. 2017 211

M. Sutton, P.L. Reu (eds.), International Digital Imaging Correlation Society, Conference Proceedings
of the Society for Experimental Mechanics Series, DOI 10.1007/978-3-319-51439-0_50
212 F. Yasmeen et al.

50.2 Experiment

Figure 50.1 shows the experimental setup. The two LED light sources illuminated the specimen at the angles shown to
minimize the potential for imaging of reflections. The two cameras were mounted firmly to the optical bench to eliminate
unwanted camera motions during the experiment. The distance from the camera lenses to the specimen was approximately
0.3048 m. The resulting speckle pattern and grey level histogram for speckle image are shown in Fig. 50.2. Table 50.1
summarizes the components in the stereovision optical system. Table 50.2 shows the parameters used to analyze the images
and acquire the displacements and strains.

Fig. 50.1 Experimental setup for out of plane rotation (a) full experimental setup, (b) camera and sample positions

Fig. 50.2 Speckle Pattern (a) image, (b) grey level histogram for image shown
50 Experimental Study of Measurement Errors in 3D-DIC Due to Out-of-Plane Specimen Rotation 213

Table 50.1 Optical system Stereovision system components

components
Cameras Grasshopper3 GS3-U3-91S6M (8bits, 3376X2704)
Lenses Schneider XENOPLAN 1.9/35-0511
Lighting White LED Lighting

Table 50.2 Parameters used StereoDIC analysis parameters

for analysis of images
Subset size 47  47 pixels2
Step size 13 pixels
Software VIC-3D
Strain filter size 5  5

Table 50.3 Calibration parameters for stereovision system

Parameter Cam0 – SAC Cam1 – SAC Cam0 – LAC Cam1 – LAC
Cx(pixel) 1676.743 1712.878 1656.124 1728.674
Cy(pixel) 1372.491 1432.081 1359.785 1430.952
fx (pixel) 11,176.322 11,134.508 11,140.056 11,114.127
fy (pixel) 11,175.219 11,133.544 11,138.959 11,112.591
Skew 3.742 2.538 1.635 0.958
›1 0.273 0.225 0.271 0.223
›2 0 0 0 0
›3 0 0 0 0

Table 50.4 Camera 0!1 transformation parameter for stereovision system

Camera 0!1 transformation Small angle calibration Large angle calibration
Alpha(ı ) 0:288ı 0:341ı
ı ı
Beta( ) 15:966 15:815ı
Gamma(ı ) 1:489ı 1:482ı
Tx (mm) 77:473 77:412
Ty (mm) 0:992 1:006
Tz (mm) 10:229 10:487
Baseline(mm) 78:1521 78:1256

To determine the effect of calibration procedures on the accuracy of stereovision system measurements in the presence
of large rotations, the investigators performed a set of baseline experiments with two calibrations. Designating the first
calibration process as the “Small Angle Calibration” (SAC), a total of 61 images of a translated and rotated planar dot pattern
[4, 5] were acquired where the rotation angles in all directions were less than 10ı . Designating the second calibration process
as the “Large Angle Calibration” (LAC) process, a total of 142 images are used, where the first 61 images used for LAC are
supplemented with 81 images as the same grid is rotated out of plane in 5ı increments within the range 40ı  ™ 40ı . Tables
50.3 and 50.4 show the calibration parameters obtained using the SAC and LAC process described above. The parameters
shown include: image plane center location, (Cx, Cy), in pixels for each camera; the factors (fx , fy ) in pixel for each camera
in the horizontal and vertical directions; skew, the deviation from orthogonality between the row and column directions in
the sensor plane; radial distortion coefficient ›1 , ›2 , ›3 ;position of pinhole in camera 1 relative to camera 0 (tx , ty , tz )(mm);
relative orientation of camera 1 with respect to camera 0 (’, “, ”) (degrees), where ’ is the relative tilt, “ the relative pan
angle, and ” the swing angle.

50.3 Experimental Results

When using the SAC process, Fig. 50.3 shows the averaged surface strains "xx . "yy and "xy that are obtained in the central 1/3
portion of the specimen as rigid body out of plane rotation is applied. Though all the strains should be zero, only the strain
"yy is near zero for all rotation angles. For the other in-plane strains, as shown in Fig. 50.3 the measured average normal
214 F. Yasmeen et al.

Fig. 50.3 Measured average in-plane strain as a function of out-of-plane rotation angle using the Small Angle Calibration process when specimen
undergoes only rigid body out of plane rotation

Fig. 50.4 Measured average "xx using both Small Angle Calibration and Large Angle Calibration procedures when specimen undergoes only
rigid body out of plane rotation

strain "xx increases in a highly non-linear manner with rotation angle, reaching a maximum value 1400 ". In addition, the
in-plane shear strain "xy is small but increases linearly with rotation angle.
Since the largest strain errors occur in the x-direction, Fig. 50.4 directly compares the experimental results for "xx obtained
when using SAC and LAC procedures. As shown in Fig. 50.4, the use of several calibration grid images that are rotated out
of plane in the range 40ı  ™  40ı provides significant improvement in the measured "xx strain field.
Inspection of the camera parameters in Tables 50.3 and 50.4 shows that the changes in camera parameters are quite
small, with all but skew changing in the 3rd significant figure or later. The results clearly demonstrate the importance of the
calibration process for the accuracy of strain measurements obtained by stereo image correlation. It is noted that, in separate
experiments, it was confirmed that mere out-of-plane translations of the calibration target do not sufficiently improve results,
while acquiring images of the grid undergoing large rotations provides significant improvement.
50 Experimental Study of Measurement Errors in 3D-DIC Due to Out-of-Plane Specimen Rotation 215

50.4 Outlook and Conclusions

ı
The influence of out of plane rotation in the range 40  ™ 40ı on strain measurements was investigated. The results
illustrate that careful attention must be paid to the calibration process used for stereo DIC systems. Using only small rotations
of the calibration target, errors in strain in excess of 1400 " were obtained. Using large target rotations, the errors could
be suppressed to around 200 " in the experiments reported here. In our attempt to further suppress this bias, a promising
calibration scheme using a speckle pattern for calibration has shown great potential, with bias levels around 50 ". The
speckle-based calibration process will be the focus of further investigation.

References

1. Schreier, H., Orteu, J.J., Sutton, M.A.: Image Correlation for Shape, Motion and Deformation Measurements. Springer (2009)
2. Sutton, M.A., Wolters, W.J., Peters, W.H., Ranson, W.F., McNeill, S.R.: Determination of displacements using an improved digital correlation
method. Image Vis. Comput. 1(3), 133–139 (1983)
3. Kahn-Jetter, Z.L., Chu, T.C.: Three-dimensional displacement measurements using digital image correlation and photogrammic analysis. Exp.
Mech. 30(1), 10–16 (1990)
4. Luo, P.F., Chao, Y.J., Sutton, M.A., Peters Iii, W.H.: Accurate measurement of three-dimensional deformations in deformable and rigid bodies
using computer vision. Exp. Mech. 33(2), 123–132 (1993)
5. Luo, P.F., Chao, Y.J., Sutton, M.A.: Application of stereo vision to three-dimensional deformation analyses in fracture experiments. Opt. Eng.
33(3), 981–990 (1994)
6. Sutton, M.A., Yan, J.H., Tiwari, V., Schreier, H.W., Orteu, J.J.: The effect of out-of-plane motion on 2D and 3D digital image correlation
measurements. Opt. Lasers Eng. 46(10), 746–757 (2008)